The widespread adoption of automated teller machines in the 1980s prompted considerable anxiety about the future of bank tellers, a scenario that holds surprising parallels to current discussions surrounding artificial intelligence and its potential disruption of the labor market. While many predicted mass displacement, the reality proved more nuanced. Initially, the number of tellers required at individual bank branches did see some reduction as ATMs handled routine transactions. However, this automation also lowered operational costs, enabling banks to open a greater number of branches. Consequently, the total employment of U.S. bank tellers remained largely consistent between 1980 and 2010. For those working in the banking sector during that period, the ATM represented less of an existential threat than contemporary forecasts suggested.
This period also saw an expansion of retail banking that created new roles. Beyond the tellers, banks began to hire more loan officers, credit analysts, personal bankers, and specialists in fraud and risk assessment. The nature of work within bank branches shifted, moving away from simple transaction processing towards more complex customer relationship management. This evolution underscores a critical point: isolated task automation, as exemplified by the early ATM, rarely leads to widespread job losses unless an occupation is built around an extremely narrow set of activities. Few switchboard operators remain, for instance, but most professions adapt.
A more significant shift in banking employment arrived around 2010 with the advent of mobile banking. This technology differed fundamentally from the ATM. While ATMs automated a task, mobile banking effectively automated the entire journey to a bank branch for many everyday activities. By 2025, merely 9% of bank customers identified branches as their primary banking channel, a stark contrast to the 36% reported in 2007. This transformation subsequently led to a decline in bank teller employment. It is important to note that this change was not solely driven by technological advancement. The Electronic Signatures in Global and National Commerce Act of 2000, which granted electronic signatures the same legal standing as ink signatures, played a crucial role in enabling fully digital banking experiences and accelerating the move away from in-person transactions.
The disruption introduced by mobile banking also generated entirely new categories of employment, including cybersecurity analysts, digital product managers, payment-platform engineers, and data-platform operators. This historical progression offers a valuable perspective on the current discourse surrounding AI. If AI follows the trajectory of other general-purpose technologies like electricity or personal computers, it is likely to enable the creation of products, services, and entire industries that are currently unforeseen. This suggests that widespread fears of mass job displacement may be overstated.
Despite the concerns voiced since the arrival of large language models like ChatGPT in late 2022 regarding the potential elimination of white-collar jobs, the labor market data tells a different story. Occupations identified as having high exposure to AI have not experienced widespread employment declines. In fact, employment growth in these highly exposed sectors has generally kept pace with, or even surpassed, that of less exposed occupations. Layoff rates remain low, and while hiring has slowed, this slowdown has been broad-based rather than concentrated in AI-intensive fields.
The current state of large language models may be akin to the ATMs of the 1980s. They are powerful tools capable of automating specific tasks, but they also augment many more, enhancing worker productivity without fundamentally altering the broader structure of work. More profound labor market disruption, akin to the transition from ATMs to mobile banking, would likely require a deeper reconfiguration of business processes, organizational structures, and customer interactions that fundamentally reshapes the role of workers rather than simply removing them from the equation. The historical record of technological change indicates that capabilities alone rarely dictate employment outcomes; rather, it is how organizations redesign work around these capabilities that truly matters. According to Vanguard Chief Economist, the evidence suggests we are currently closer to the ATM phase of technological integration than the more transformative mobile banking phase.
